Trends In Autism Spectrum Disorder Prediction Using Machine Learning : A Review

Abstract

Autism Spectrum Disorder (ASD) is a neurodevelopmental disorder that significantly affects social, linguistic, and cognitive skills. Early diagnosis is crucial for improving long-term outcomes, yet traditional diagnostic methods are time-consuming and expensive. This review aims to explore the potential of machine learning techniques in enhancing the accuracy and efficiency of ASD prediction and diagnosis. By examining ten studies, the review evaluates the various machine learning (ML) algorithms used, pre-processing techniques employed, and datasets analysed. Key findings indicate that pre-processing techniques such as handling missing values, normalization, and feature selection are vital for improving model accuracy. Support Vector Machine and Logistic Regression consistently demonstrated high accuracy in predicting ASD across various datasets. The conclusion underscores the importance of pre-processing in developing reliable machine learning models for ASD prediction and highlights the need for future research to address challenges related to data accessibility, model interpretability, and validation across diverse populations. The responsible integration of ML technologies into clinical practice could revolutionize early diagnosis and intervention strategies for ASD.